Challenges in diagnosis and management of traumatic diaphragmatic injuries: experiences from a regional New Zealand trauma centre

Traumatic diaphragmatic injury (TDI) is a rare yet serious consequence of both blunt and penetrating thoraco-abdominal trauma.¹ The reported incidence varies significantly depending on the mechanism and anatomical location, with some series reporting rates from less than 1% to 8%.^2,3 The diaphragm is the primary muscle responsible for respiration; substantial defects in the diaphragm can lead to herniation of visceral organs and may result in complications such as obstruction, ischaemia and perforation.⁴ During normal respiration, a pressure difference is created between the abdomen and thorax. Consequently, diaphragmatic injuries will not heal without surgical repair and may enlarge over time.

Methods

A retrospective review of data from the TMT trauma registry from 1 January 2014 to 31 December 2023 was conducted. The TMT trauma registry includes trauma admissions of all severities across six hospitals in the region and employs Abbreviated Injury Scale (AIS) coding for injury severity and diagnosis.¹⁰ The variables collected from the registry include patient demographic details, injury event information, AIS diagnoses of TDI, body regions injured and hospital length of stay (LOS). TDI cases were identified using AIS diagnosis codes 440602, 440604, 440606, 440608, 440610 and 440699.¹¹ To investigate the diagnostic and surgical management of patients with TDI, those admitted to Waikato Hospital were selected because it is the region’s Level 1 trauma centre, accredited by the Royal Australasian College of Surgeons Trauma Care Verification Program. The clinical records were requested through Waikato Hospital’s medical records department and individually reviewed for operation notes, radiology reports and clinical details. Data on the use of chest X-ray (CXR), computed tomography (CT), indications for surgery, operative approach (modality and access) and complications were collected. Ethical clearance was deemed unnecessary by the New Zealand Health and Disability Ethics Committee (58771) as this was a retrospective review of secondary data and an audit of records intended to inform and enhance clinical care. Locality approval was obtained from the Waikato Hospital research office (RD025047).

Descriptive analysis was performed and reported using numbers, percentages, median and interquartile range (IQR). Burden of TDIs was calculated as a percentage of total trauma admissions in the TMT Region. For the statistical analysis, age was categorised into three groups: 0–14, 15–64, and 65+ years; gender into male and female; and ethnicity into Māori and non-Māori. Injury severity was classified using the Injury Severity Score (ISS) into major (ISS>12)¹² and non-major (ISS<13). Descriptive statistics on diagnosis modality, indications for surgery, surgical management and patient outcomes were generated to analyse TDI detection rates. Rate ratios for detection rates were calculated with a 95% confidence level. Complications were categorised using the Clavien-Dindo grading system.¹³ This system classifies complications from I to IV (Table 1), whereby grade I complications required minor deviations from the planned post-operative course and grade IV indicated the death of a patient resulting from surgical complications. Mortality was defined as death occurring during hospital admission, including in the emergency department (ED). Hospital LOS was calculated as the total number of days spent in the hospital at the event level.

Results

There were 56,113 trauma patients who presented to TMT hospitals from 2014 to 2023, of which 130 (0.2%) sustained a TDI (Figure 1). Eighty-eight were males (68%). The mean age was 42 years. The majority were major trauma patients with ISS>12 (96%), with a mean ISS score of 26. Nearly two-thirds (64%) were caused by blunt trauma.

View Figure 1–3, Table 1–5.

Forty-three patients were admitted to Waikato Hospital (Level 1 trauma centre) with TDI. The demographic and injury characteristics of these cases are summarised in Table 1.

The left hemidiaphragm was injured more frequently (70% left vs 30% right), and on both sides, penetrating injuries accounted for a third of the cases (33% left and 31% right). The most common injury was diaphragm laceration (<10cm) (44%), with diaphragmatic rupture (diaphragm laceration>10cm; with significant tissue loss and diaphragm rupture with herniation) occurring in 21 patients (48.8%). There were no isolated diaphragmatic injuries, and 40 patients (93%) had injuries to other body regions. Apart from the chest, the most commonly associated injured region was the abdomen, seen in 81% of cases (Table 2).

The diagnosis of diaphragmatic injury was made pre-operatively in 29 of 43 patients (67.4%). The injury was apparent on CXR in 19 patients (44%) and on CT alone in 26 patients (81%) (Figure 2). In 14 patients, the diagnosis was only confirmed intraoperatively (32.5%). The detection rate with radiological methods is significantly lower than intraoperative detection (rate ratios for CXR and CT are 0.44 and 0.81, respectively, p<0.05) (Table 3).

At the presentation, 29 patients required immediate surgery for life-threatening injuries (Figure 3). Diaphragmatic rupture was the primary indication in only eight of these 29 patients. There were three diaphragmatic ruptures in the intraoperative diagnosis group. Other indications for surgery within this group included haemodynamic instability (11); perforation (six); peritonitis (one) and other injuries requiring immediate surgery (three). After the initial resuscitation and assessment, 14 patients did not meet the indications for immediate surgery.

The timing of subsequent repair of the diaphragmatic injuries in the group that did not require immediate surgery: Seven were repaired within 24 hours and another seven were repaired beyond 24 hours (range: 2 to 4 days). In total, 36 (83.7%) had repair immediately or within 24 hours.

Thirty-four patients underwent surgery through a straight-to-open approach (79%), while minimally invasive methods were employed in 9 patients. Three patients (7%) required conversion to open surgery. Diaphragmatic injuries were predominantly repaired via abdominal access rather than thoracic (31 versus six), with a combined approach observed in six cases (Table 4). Five patients attempted initial non-operative management (11.6%). Almost all injuries were primarily repaired with sutures (95%), with mesh used in only one patient acutely when the defect could not be approximated due to tissue loss.

Overall, the complication rate for the TDI cases was 44.2% (Table 5), more common among those with penetrating injury (64.3% compared to 34.5% among those with sustained blunt injury). These complications occurred in cases with major injuries (ISS>12) with a median AIS score of 3.5 (IQR: 3–4). The most common type was Clavien-Dindo grade II (42% overall, 40% among those with blunt injury and 44% among those with penetrating injury), including pneumonia (five cases), urinary tract infections (two cases) and portal vein thrombus (one case). Five patients required re-intervention: empyema (two), abdominal wall collection/complication (two) and bile leak (one). Two patients developed multi-organ failure secondary to sepsis caused by delayed colonic perforation (one) and enterocutaneous fistula formation (one). There were two mortalities: one patient died in the emergency department following a pulseless electrical activity (PEA) arrest, and the other died of multi-organ failure. The median hospital LOS was 12 days (min 1, max 160) for the whole TDI cohort.

Discussion

Diaphragmatic injury remains an uncommon diagnosis as it is often presented with other injuries, most commonly to the chest or abdomen. The sensitivity for CXR alone was 44%; this increased to 67% with the utilisation of CT assessment. Despite significant advancements in pre-operative imaging, our series showed that diagnosis was still only made intraoperatively in approximately one-third of cases. The majority of these patients required immediate surgery, with diaphragmatic rupture being a primary indication for 27.5% of cases, alongside other critical interventions. Repairs occurred mostly within 24 hours.

The combined pre-operative detection rate of 67% in our study is comparable to those reported in other studies.¹⁴ A 2023 meta-analysis showed a sensitivity of 74%, but it can range from 56% to 83%.¹⁵ Although CT is now the gold standard for non-invasive diagnosis, the lower detection rate reflects real-world use of pre-operative imaging, with noted limitations of CT utility in penetrating diaphragmatic injuries¹ and in haemodynamically unstable patients who proceed directly to the operating theatre. In this study, nearly two-thirds of patients with TDI met the criteria for urgent surgery. To the best of our knowledge, this rate has not been specifically reported before in the literature. It suggests that diaphragmatic injury is not only associated with multisystem trauma but also serves as a surrogate marker of high-energy transfer.¹⁶ Consequently, careful intraoperative inspection for other injuries is essential when a diaphragmatic injury is identified.

Laparotomy remains the most commonly used method for operative repair in this group. Other authors have noted that laparotomy is the preferred approach.³ Our findings support this, as the abdomen was the most frequently injured area. The use of minimally invasive techniques is increasingly accepted, especially in haemodynamically stable patients. Studies indicate that laparoscopy is more often utilised for left-sided TDIs and video-thoracoscopy for right-sided TDIs.¹⁷ The optimal surgical approach for traumatic diaphragmatic hernias has not been definitively established. However, current guidelines endorse the use of laparoscopy in stable patients, both to confirm the diagnosis and to perform the repair.^1,8,18 The choice depends on multiple factors, including the patient’s haemodynamic status, suspected associated injuries and available local resources.

The mortality rate for cases with TDI is lower than some estimates of TDI-related deaths in the literature;³ however, mortality does not accurately reflect TDI-related death, as it is affected by associated injuries. The true rate is difficult to determine with lower-grade diaphragmatic injuries that are asymptomatic and undetectable. Although some TDIs are only identified during autopsy,¹⁹ some experts believe that part of the variability could be due to the increasing use of CT imaging in trauma patients, which has led to an overall rise in the incidence of these injuries over time.²⁰

Clavien-Dindo grade II complications have been recognised as the most common in trauma cases,²¹ as observed in this cohort of TDI. The rate of complications, classified as Clavien-Dindo grade III or higher, is similar to that found in the Pietermaritzburg Metropolitan Trauma Service (PMTS) study, which examined complications related to TDI repair.²² However, the complications observed in this cohort may be affected by the overall injury syndrome, of which diaphragmatic injury is a component. The PMTS study noted that 15% of patients required surgical interventions due to pulmonary lobar collapse, bile leak, abdominal evisceration, iatrogenic injury, anastomotic leak, intra-abdominal abscess or a missed injury.²²

The retrospective nature of the study design is the primary limitation. Despite the comprehensive data collection in the TMT trauma registry, it is not linked to long-term follow-up, which restricts the ability to detect delayed presentations of traumatic diaphragmatic hernia. Future prospective studies may clarify clinical factors that predict missed diagnosis rates and delayed presentations by type of TDI. While there may be some limitations in the use of AIS, it is consistent with the American Association for the Surgery of Trauma (AAST) classification into five grades. Future studies may consider using the AAST classification alongside AIS and examining the associations with the mechanism of injury. Furthermore, as all TDI cases had associated injuries, outcomes cannot be directly attributed to TDI. Scoping the detailed investigation to only admissions at the Level 1 trauma centre limits its generalisation to the region; nevertheless, this study presents the current challenges in the diagnosis and management, which can be built upon in wider studies on TDI in the region. 

Conclusion

Pre-operative imaging alone has limitations in identifying TDI requiring repair. Our research showed a low threshold for operative intervention in patients with suspected diaphragmatic injuries who were haemodynamically unstable. A high index of suspicion of TDI is necessary, especially in unstable patients with injury syndromes.